Increase exposure by publishing ZARP on tool/workflow registries
uniqueg opened this issue · comments
Is your feature request related to a problem? Please describe.
WorkflowHub is a FAIR registry for scientific computational workflows available on the European Open Science Cloud (EOSC) marketplace.
While still relatively new, WorkflowHub is a growing resource that is heavily supported by international agencies such as ELIXIR, and thus it is likely to draw increasing audiences from the life science community over the coming years. The service also has features that make it easy to discover/fetch - via implementing the GA4GH Tool Registry Service (TRS) API - and launch workflows in the cloud, allowing researchers to more easily test or use them.
Among other workflow types, WorkflowHub supports publishing Snakemake workflows, and indeed, more than a dozen Snakemake workflows are already available on it.
Describe the solution you'd like
A clear and concise description of what you want to happen.
Publish ZARP on WorkflowHub and, ideally, implement an automatic update whenever a new release is pushed (building on #87).
Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.
A similar, more established workflow registry popular in the North American market is Dockstore. Previously, Snakemake workflows were not supported, however, this may have changed. Still, there currently don't seem to be any Snakemake workflows in the registry. Trying to publish ZARP on Dockstore could be another way of further exposing our workflow; however, it may be more difficult given that there are no examples to build on.
Yet another registry for any kind of life science software (not specific to but also including workflows) is bio.tools, which is also heavily supported by ELIXIR and features more than 25'000 tools.
Additional context
Add any other context or screenshots about the feature request here.
@DrYak (located in Basel) could possibly help us with the registration and/or automating publication, as their V-PIPE Snakemake workflow is available on WorkflowHub.
Status: I created an account for Zavolab at Workflowhub on Monday, but still waiting for them to approve our request to have a "Team" (which we need in order to upload a workflow)
Status: registered team "Zavolan Lab", registered workflow "ZARP" via github.
Open Questions @uniqueg :
- License correct (Apache Software License 2.0)?
- Discussion Channels? Is it possible to create a public slack channel? I put github issues for now
- Our main branch is linked, so we have to make sure it is up to date
- is the maturity "stable" already? Other option is "work-in-progress". I put "Not specified" for now
- suggestions for tags, other than "bioinformatics", "rna", "rna-seq"?
- couldn't get the images (logo and schema) render in the markdown description
- couldn't get any of the links to work in the markdown description
Thanks @ninsch3000 :)
My 2 cents on your questions:
- Yes, license is okay
- For discussion channels, I think it is actually possible to open something like that on GitHub. It's called "Discussions" or something. Don't remember if this can be enabled individually for each repo or only per organization. If the former works, I think this would be a good option.
- Yes, makes sense to put
main - Yes, leave "Not specified" for now and set to "Mature" upon publication
- Hmm, not really, maybe "ngs" and/or "high throughout"
- I don't know what to say about the rendering issues. I can ask in the WorkflowHub Slack. Did you try non-bibliography style links, i.e.
[]()instead of[][]? And what's the issue with the schema? I see one in the screenshot you attached...
So we have a Discussion forum activated as well, with a first "Welcome" post: https://github.com/zavolanlab/zarp/discussions
@ninsch3000 Any other updates on this? As far as I'm concerned, if Workflow Hub is done and the issues mentioned above are resolved, I think we can close this issue.
- added GitHub Discussions as discussion channel
- added "ngs" and "high throughput" tags
- what is the doi DOI:10.1101 we provide in the github badge? Workflowhub also provides the generation of a DOI, but I suppose we don't need another one?
- added links manually, should work now.
Awesome, thanks :)
If WorkflowHub mints a DOI for us, then I agree that we don't need to add another one.
So I think you can close the issue
Sorry, have to re-phrase my question: Do we want Workflowhub to create a DOI? If so, would we want to delete the DOI that is shown in the github badge in our repo? Where does that one come from?
I see... https://doi.org/10.1101/2021.11.18.469017 is from biorxiv, 10.5281/zenodo.5703358 is from zenodo.
So do we additionally want Workflowhub to create a DOI?
I'd say yes. And maybe instead of Zenodo. Better to use more specialized registries when available. I don't think there's much added value of keeping it up-to-date on Zenodo. Once we have the WFH, we can reference it in Zenodo and say that the Zenodo record will not be updated.
But we can also keep both, or just Zenodo. I don't have a strong opinion here.
Ok, here we have a Workflowhub DOI: https://doi.org/10.48546/WORKFLOWHUB.WORKFLOW.447.1
Awesome. Do you wanna create a shield/badge in the README.md for this one as well?
I would say that for now we can just leave all three in. Perhaps you could somehow manage to change the name of the shields so that one says Zenodo, the other bioRxiv and the new one WorkflowHub - and all with the same blue color? I think this should work somehow via shields.io. But if it's too much of a hassle, just add the new one with a different color.
And then close this issue with the PR :)